This thesis evaluates the ability of covalent linkages to induce protein secondary structures in very short cyclic peptides. A combination 1D and 2D 1H NMR spectroscopy, circular dichroism spectroscopy and molecular dynamics simulations were employed to elucidate the three dimensional structures (in water) of cyclic peptides synthesized by different linkage strategies. Ramachandran analysis and molecular modelling were subsequently used to compare peptide backbone geometries with protein secondary structure units. Alpha-helical, α-turn, pseudo-helical, β-sheet and β-strand peptidomimetics were synthesized and characterized. Factors controlling the backbone geometry of small cyclic peptides are discussed and a methodology for the rational design of protein secondary structures in short peptides through selection of the appropriate covalent linkage is presented.
The aim of Chapter 2 was to construct the shortest possible α-helix. The shortest reported α -helical peptide, Ac-(cyclo-1,5)-[KAAAD]-NH2 (Chapter 5) contains 5 amino acids, although only 3.6 amino acid residues are theoretically required to form a single α-helical alpha turn. The N-termini of tripeptides H2N-ALE-NH2 and H2N-ALhE-NH2 were constrained by a covalent linkage between the side-chain of the (i+2)nd residue and N-terminus of the ith residue as it was hoped that the linkage would constrain (cyclo-1,3)-[ALE]-NH2 and (cyclo-1,3)-[ALhE]-NH2 to a single α-helical turn. Cyclic peptides were investigated intensively by 1H NMR, CD spectroscopy and MD simulation. Instead of a single α-helical alpha turn, (cyclo-1,3)-[ALE]-NH2 was found to be a highly constrained cyclic peptide featuring two antiparallel β-strands connected by a cis-amide bond. Homologue (cyclo-1,3)-[ALhE]-NH2 was found to be a β-III type turn.
The aim of Chapter 3 was to isolate and characterize the first peptide mimetics of “non-classical” α-turns. Non-classical α-turns are structurally distinct from α-helical alpha turns (in analogy to the family of β-turn types) and are found at helix termini and at kinks near the active sites of many proteins. A series of tetrapeptides were synthesized by linking the side-chain of residue i to the C-terminus of residue i+3 (e.g. Ac-(cyclo-1,4)-[DapARA]) and the side-chain of residue i+3 to the N-terminus of residue i (e.g. (cyclo-1,4)-[ARAE]-NH2) and intensively evaluated by 1H NMR spectroscopy and CD spectroscopy and molecular dynamics simulations. The ith, (i+1)st and (i+2)nd residues of two peptides (vide supra) were found to match the Ramachandran space of specific types of non-classical α-turns (I-α RS and II-α LU) and also overlay with the peptide backbone of non-classical alpha turns in protein crystal structures.
The aim of Chapter 4 was to isolate and characterize additional non-classical α-turn types. A series of pentapeptides was synthesized by linking the side-chain of residue i to the C-terminus of residue i+4 (e.g. Ac-(cyclo-1,5)-[KARAL]). Peptides were intensively investigated by 1H NMR, CD spectroscopy and molecular dynamics simulations; a NMR structure was calculated for each peptide and represented in Ramachandran space. No additional non-classical α-turn types were discovered. The peptide backbone was found to sequentially deconvolute from a α-turn to a novel pseudo-helical geometry as the linker length was increased from one to five methylene units. The aliphatic linker, previously thought to be flexible, was found to control the peptide backbone geometry in a predicable manner.
The aim of Chapter 5 was to investigate the ability of the short α-helical peptide Ac-(cyclo-1,4)-[KARAD]-NH2 to nucleate α-helicity in a linear peptide chain. Ac-(cyclo-1,4)-[KARAD]-NH2 has both hydrogen bond accepting carbonyl groups and hydrogen bond donating amide groups and was thus predicted to nucleate α-helicity when appended to both the C-terminus and N-terminus of a peptide chain. The 1H NMR solution structure of Ac-(cyclo-1,4)-[KAAAD]-NH2 in water was calculated and used to infer the geometry of Ac-(cyclo-1,4)-[KARAD]-NH2. The comparison of the ability of Ac-(cyclo-1,4)-[KARAD]-NH2 to nucleate α -helicity in a palindromic linear peptide, when employed as a C-terminal or a N-terminal cap, was evaluated by CD and NMR spectroscopy. The induction of α-helicity by Ac-(cyclo-1,4)-[KARAD]-NH2 when appended to the C-terminus of a series of linear peptides of increasing length was evaluated by CD and NMR spectroscopy and found to be superior to induction of α -helicity from the N-terminus. Structural arguments derived from extensive 1D and 2D 1H NMR experiments are advanced to rationalize the results.
Chapter 6 provides an overview of the solution structures calculated in Chapters 2-5. Factors controlling the backbone geometry of small cyclic peptides are enumerated and the limitations of cyclization strategies in reproducing protein secondary structure units in small peptides are discussed. Lastly, work-in-progress comparing the ability of several classes of i→ (i+7) linker to induce well-defined solution structure in octapeptides is presented as an Appendix.